-
Notifications
You must be signed in to change notification settings - Fork 76
/
zipstream.py
412 lines (349 loc) · 15.2 KB
/
zipstream.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
#!/usr/bin/env python
"""
Iterable ZIP archive generator.
Derived directly from zipfile.py
"""
import struct, os, time, sys
import binascii
try:
import zlib # We may need its compression method
except ImportError:
zlib = None
__all__ = ["ZIP_STORED", "ZIP_DEFLATED", "ZipStream"]
ZIP64_LIMIT= (1 << 31) - 1
# constants for Zip file compression methods
ZIP_STORED = 0
ZIP_DEFLATED = 8
# Other ZIP compression methods not supported
# Here are some struct module formats for reading headers
structEndArchive = "<4s4H2lH" # 9 items, end of archive, 22 bytes
stringEndArchive = "PK\005\006" # magic number for end of archive record
structCentralDir = "<4s4B4HlLL5HLl"# 19 items, central directory, 46 bytes
stringCentralDir = "PK\001\002" # magic number for central directory
structFileHeader = "<4s2B4HlLL2H" # 12 items, file header record, 30 bytes
stringFileHeader = "PK\003\004" # magic number for file header
structEndArchive64Locator = "<4slql" # 4 items, locate Zip64 header, 20 bytes
stringEndArchive64Locator = "PK\x06\x07" # magic token for locator header
structEndArchive64 = "<4sqhhllqqqq" # 10 items, end of archive (Zip64), 56 bytes
stringEndArchive64 = "PK\x06\x06" # magic token for Zip64 header
stringDataDescriptor = "PK\x07\x08" # magic number for data descriptor
# indexes of entries in the central directory structure
_CD_SIGNATURE = 0
_CD_CREATE_VERSION = 1
_CD_CREATE_SYSTEM = 2
_CD_EXTRACT_VERSION = 3
_CD_EXTRACT_SYSTEM = 4 # is this meaningful?
_CD_FLAG_BITS = 5
_CD_COMPRESS_TYPE = 6
_CD_TIME = 7
_CD_DATE = 8
_CD_CRC = 9
_CD_COMPRESSED_SIZE = 10
_CD_UNCOMPRESSED_SIZE = 11
_CD_FILENAME_LENGTH = 12
_CD_EXTRA_FIELD_LENGTH = 13
_CD_COMMENT_LENGTH = 14
_CD_DISK_NUMBER_START = 15
_CD_INTERNAL_FILE_ATTRIBUTES = 16
_CD_EXTERNAL_FILE_ATTRIBUTES = 17
_CD_LOCAL_HEADER_OFFSET = 18
# indexes of entries in the local file header structure
_FH_SIGNATURE = 0
_FH_EXTRACT_VERSION = 1
_FH_EXTRACT_SYSTEM = 2 # is this meaningful?
_FH_GENERAL_PURPOSE_FLAG_BITS = 3
_FH_COMPRESSION_METHOD = 4
_FH_LAST_MOD_TIME = 5
_FH_LAST_MOD_DATE = 6
_FH_CRC = 7
_FH_COMPRESSED_SIZE = 8
_FH_UNCOMPRESSED_SIZE = 9
_FH_FILENAME_LENGTH = 10
_FH_EXTRA_FIELD_LENGTH = 11
class ZipInfo (object):
"""Class with attributes describing each file in the ZIP archive."""
__slots__ = (
'orig_filename',
'filename',
'date_time',
'compress_type',
'comment',
'extra',
'create_system',
'create_version',
'extract_version',
'reserved',
'flag_bits',
'volume',
'internal_attr',
'external_attr',
'header_offset',
'CRC',
'compress_size',
'file_size',
)
def __init__(self, filename="NoName", date_time=(1980,1,1,0,0,0)):
self.orig_filename = filename # Original file name in archive
# Terminate the file name at the first null byte. Null bytes in file
# names are used as tricks by viruses in archives.
null_byte = filename.find(chr(0))
if null_byte >= 0:
filename = filename[0:null_byte]
# This is used to ensure paths in generated ZIP files always use
# forward slashes as the directory separator, as required by the
# ZIP format specification.
if os.sep != "/" and os.sep in filename:
filename = filename.replace(os.sep, "/")
self.filename = filename # Normalized file name
self.date_time = date_time # year, month, day, hour, min, sec
# Standard values:
self.compress_type = ZIP_STORED # Type of compression for the file
self.comment = "" # Comment for each file
self.extra = "" # ZIP extra data
if sys.platform == 'win32':
self.create_system = 0 # System which created ZIP archive
else:
# Assume everything else is unix-y
self.create_system = 3 # System which created ZIP archive
self.create_version = 20 # Version which created ZIP archive
self.extract_version = 20 # Version needed to extract archive
self.reserved = 0 # Must be zero
self.flag_bits = 0x08 # ZIP flag bits, bit 3 indicates presence of data descriptor
self.volume = 0 # Volume number of file header
self.internal_attr = 0 # Internal attributes
self.external_attr = 0 # External file attributes
# Other attributes are set by class ZipFile:
# header_offset Byte offset to the file header
# CRC CRC-32 of the uncompressed file
# compress_size Size of the compressed file
# file_size Size of the uncompressed file
def DataDescriptor(self):
if self.compress_size > ZIP64_LIMIT or self.file_size > ZIP64_LIMIT:
fmt = "<4slQQ"
else:
fmt = "<4slLL"
return struct.pack(fmt, stringDataDescriptor, self.CRC, self.compress_size, self.file_size)
def FileHeader(self):
"""Return the per-file header as a string."""
dt = self.date_time
dosdate = (dt[0] - 1980) << 9 | dt[1] << 5 | dt[2]
dostime = dt[3] << 11 | dt[4] << 5 | (dt[5] // 2)
if self.flag_bits & 0x08:
# Set these to zero because we write them after the file data
CRC = compress_size = file_size = 0
else:
CRC = self.CRC
compress_size = self.compress_size
file_size = self.file_size
extra = self.extra
if file_size > ZIP64_LIMIT or compress_size > ZIP64_LIMIT:
# File is larger than what fits into a 4 byte integer,
# fall back to the ZIP64 extension
fmt = '<hhqq'
extra = extra + struct.pack(fmt,
1, struct.calcsize(fmt)-4, file_size, compress_size)
file_size = 0xffffffff # -1
compress_size = 0xffffffff # -1
self.extract_version = max(45, self.extract_version)
self.create_version = max(45, self.extract_version)
header = struct.pack(structFileHeader, stringFileHeader,
self.extract_version, self.reserved, self.flag_bits,
self.compress_type, dostime, dosdate, CRC,
compress_size, file_size,
len(self.filename), len(extra))
return header + self.filename + extra
class ZipStream:
"""
"""
def __init__(self, path, arc_path='', compression=ZIP_DEFLATED):
if compression == ZIP_STORED:
pass
elif compression == ZIP_DEFLATED:
if not zlib:
raise RuntimeError,\
"Compression requires the (missing) zlib module"
else:
raise RuntimeError, "That compression method is not supported"
self.filelist = [] # List of ZipInfo instances for archive
self.compression = compression # Method of compression
self.path = path # source path
self.arc_path = arc_path # top level path in archive
self.data_ptr = 0 # Keep track of location inside archive
def __iter__(self):
for data in self.zip_path(self.path, self.arc_path):
yield data
yield self.archive_footer()
def update_data_ptr(self, data):
"""As data is added to the archive, update a pointer so we can determine
the location of various structures as they are generated.
data -- data to be added to archive
Returns data
"""
self.data_ptr += len(data)
return data
def zip_path(self, path, archive_dir_name):
"""Recursively generate data to add directory tree or file pointed to by
path to the archive. Results in archive containing
archive_dir_name/basename(path)
archive_dir_name/basename(path)/*
archive_dir_name/basename(path)/*/*
.
.
.
path -- path to file or directory
archive_dir_name -- name of containing directory in archive
"""
if os.path.isdir(path):
dir_name = os.path.basename(path)
for name in os.listdir(path):
r_path = os.path.join(path, name)
r_archive_dir_name = os.path.join(archive_dir_name, dir_name)
for data in self.zip_path(r_path, r_archive_dir_name):
yield data
else:
archive_path = os.path.join(archive_dir_name, os.path.basename(path))
for data in self.zip_file(path, archive_path):
yield data
def zip_file(self, filename, arcname=None, compress_type=None):
"""Generates data to add file at 'filename' to an archive.
filename -- path to file to add to arcive
arcname -- path of file inside the archive
compress_type -- unused in ZipStream, just use self.compression
This function generates the data corresponding to the fields:
[local file header n]
[file data n]
[data descriptor n]
as described in section V. of the PKZIP Application Note:
http://www.pkware.com/business_and_developers/developer/appnote/
"""
st = os.stat(filename)
mtime = time.localtime(st.st_mtime)
date_time = mtime[0:6]
# Create ZipInfo instance to store file information
if arcname is None:
arcname = filename
arcname = os.path.normpath(os.path.splitdrive(arcname)[1])
while arcname[0] in (os.sep, os.altsep):
arcname = arcname[1:]
zinfo = ZipInfo(arcname, date_time)
zinfo.external_attr = (st[0] & 0xFFFF) << 16L # Unix attributes
if compress_type is None:
zinfo.compress_type = self.compression
else:
zinfo.compress_type = compress_type
zinfo.file_size = st.st_size
zinfo.header_offset = self.data_ptr # Start of header bytes
fp = open(filename, "rb")
zinfo.CRC = CRC = 0
zinfo.compress_size = compress_size = 0
zinfo.file_size = file_size = 0
yield self.update_data_ptr(zinfo.FileHeader())
if zinfo.compress_type == ZIP_DEFLATED:
cmpr = zlib.compressobj(zlib.Z_DEFAULT_COMPRESSION,
zlib.DEFLATED, -15)
else:
cmpr = None
while 1:
buf = fp.read(1024 * 8)
if not buf:
break
file_size = file_size + len(buf)
CRC = binascii.crc32(buf, CRC)
if cmpr:
buf = cmpr.compress(buf)
compress_size = compress_size + len(buf)
yield self.update_data_ptr(buf)
fp.close()
if cmpr:
buf = cmpr.flush()
compress_size = compress_size + len(buf)
yield self.update_data_ptr(buf)
zinfo.compress_size = compress_size
else:
zinfo.compress_size = file_size
zinfo.CRC = CRC
zinfo.file_size = file_size
yield self.update_data_ptr(zinfo.DataDescriptor())
self.filelist.append(zinfo)
def archive_footer(self):
"""Returns data to finish off an archive based on the files already
added via zip_file(...). The data returned corresponds to the fields:
[archive decryption header]
[archive extra data record]
[central directory]
[zip64 end of central directory record]
[zip64 end of central directory locator]
[end of central directory record]
as described in section V. of the PKZIP Application Note:
http://www.pkware.com/business_and_developers/developer/appnote/
"""
data = []
count = 0
pos1 = self.data_ptr
for zinfo in self.filelist: # write central directory
count = count + 1
dt = zinfo.date_time
dosdate = (dt[0] - 1980) << 9 | dt[1] << 5 | dt[2]
dostime = dt[3] << 11 | dt[4] << 5 | (dt[5] // 2)
extra = []
if zinfo.file_size > ZIP64_LIMIT or zinfo.compress_size > ZIP64_LIMIT:
extra.append(zinfo.file_size)
extra.append(zinfo.compress_size)
file_size = 0xffffffff #-1
compress_size = 0xffffffff #-1
else:
file_size = zinfo.file_size
compress_size = zinfo.compress_size
if zinfo.header_offset > ZIP64_LIMIT:
extra.append(zinfo.header_offset)
header_offset = -1 # struct "l" format: 32 one bits
else:
header_offset = zinfo.header_offset
extra_data = zinfo.extra
if extra:
# Append a ZIP64 field to the extra's
extra_data = struct.pack('<hh' + 'q'*len(extra),1, 8*len(extra), *extra) + extra_data
extract_version = max(45, zinfo.extract_version)
create_version = max(45, zinfo.create_version)
else:
extract_version = zinfo.extract_version
create_version = zinfo.create_version
centdir = struct.pack(structCentralDir,
stringCentralDir, create_version,
zinfo.create_system, extract_version, zinfo.reserved,
zinfo.flag_bits, zinfo.compress_type, dostime, dosdate,
zinfo.CRC, compress_size, file_size,
len(zinfo.filename), len(extra_data), len(zinfo.comment),
0, zinfo.internal_attr, zinfo.external_attr,
header_offset)
data.append( self.update_data_ptr(centdir))
data.append( self.update_data_ptr(zinfo.filename))
data.append( self.update_data_ptr(extra_data))
data.append( self.update_data_ptr(zinfo.comment))
pos2 = self.data_ptr
# Write end-of-zip-archive record
if pos1 > ZIP64_LIMIT:
# Need to write the ZIP64 end-of-archive records
zip64endrec = struct.pack(structEndArchive64, stringEndArchive64,
44, 45, 45, 0, 0, count, count, pos2 - pos1, pos1)
data.append( self.update_data_ptr(zip64endrec))
zip64locrec = struct.pack(structEndArchive64Locator,
stringEndArchive64Locator, 0, pos2, 1)
data.append( self.update_data_ptr(zip64locrec))
# XXX Why is `pos3` computed next? It's never referenced.
pos3 = self.data_ptr
endrec = struct.pack(structEndArchive, stringEndArchive,
0, 0, count, count, pos2 - pos1, -1, 0)
data.append( self.update_data_ptr(endrec))
else:
endrec = struct.pack(structEndArchive, stringEndArchive,
0, 0, count, count, pos2 - pos1, pos1, 0)
data.append( self.update_data_ptr(endrec))
return ''.join(data)
if __name__ == "__main__":
zipfile = sys.argv[1]
path = sys.argv[2]
zf = open(zipfile, 'wb')
for data in ZipStream(path):
zf.write(data)
zf.close()